GLS 2005 DEM - 磨房

billzeng OP • 2011-02-03 09:04

同一地区几种DEM的对比

Dyson(丁丁) • 2011-02-05 05:40

下了一个数据，台湾最南端同一地点进行比较：

Srtm4.1，20M

Aster，20M

GLS2005，20M

yangl1996 • 2011-02-06 07:58

看来我该放弃strm了

工兵pro • 2011-02-07 15:50

大家老是用缩写，怪不得一上来找不到
原来GLS是美国地质调查局全球陆地测绘（Global Land Survey）系列数据，2005是最新的一个

Product Description
In the past, the U.S. Geological Survey (USGS) and the National Aeronautics and Space Administration (NASA) collaborated on the creation of three global land datasets using Landsat data: one from the 1970s, and one each from circa 1990 and 2000. Each of these global data sets was created from the primary Landsat sensor in use at the time: the Multispectral Scanner in the 1970s, the Thematic Mapper (TM) in 1990s, and the Enhanced Thematic Mapper Plus (ETM+) in 2000.

USGS and NASA collaboration have extended the Global Landsat Data collection with the addition of core acquisition dates from 2005 to 2006. The data includes both Landsat TM and ETM+ imagery as well as EO-1 ALI imagery over islands, making GLS the first ever global dataset built with data from two sensors.

Direct downlink to a ground station is the only way to acquire Landsat 5 TM data, because unlike Landsat 7, the Landsat 5 satellite does not have an onboard solid-state recorder. To obtain Landsat 5 TM imagery for areas outside the United States, the USGS negotiated agreements with ground stations around the globe to downlink and send data to the USGS Landsat Ground Station. These agreements make thousands of additional Landsat 5 TM scenes available for GLS consideration.

The USGS Landsat Ground Station archives nearly 100,000 new Landsat 7 ETM+ images every year. These images are a record of nearly all the land area on Earth. The Landsat 7 ETM+ and Landsat 5 TM data incorporated into the GLS dataset must meet quality and cloud cover standards. Data recorded in 2004 and 2007 will be used only as needed to fill areas of low image quality or excessive cloud cover.

The GLS dataset scene-based, Individual scenes are made available as they are processed. EO-1 Level 1 geometrically corrected ALI data were also used to fill gaps in the GLS 2005 dataset.

http://landsat.usgs.gov/index.php

The following data products are available from the Global Land Survey data collection:

GLS 2005
GLS 2005 - Islands (EO-1)
GLS 2000
GLS 1990
GLS 1975
The acquisition dates for the GLS 1975 July 25, 1972 to July 13, 1987.

The acquisition dates for the GLS 1990 is April 1984 to November 1997.

All scenes were acquired between 2004 and 2007. The GLS 2000 data was used to correct this dataset.

The acquisition dates for the GLS 2005 is 2005 (+/- 3 years).

GLS 2005 will continue to grow as we complete this collection.

数据指标是30m分辨率，和ASTER一样，似乎主要由红外成像获取，同样会缺乏SRTM的云层，雨雾，植被穿透能力

USGS提到重采样方法是Cubic Convolution （立方卷积），这也许是它优于SRTM 线性插值的地方，SRTM在从原始30米分辨率降到90米时用的算法实在太二了，严重损失了高频分量（对应小地形分辨力）

大家先用熟悉的山区比较比较
等高线地图的效果和很多因素有关，billzeng的样图是什么经纬度的，我也做一个比较比较

Dyson(丁丁) • 2011-02-08 01:53

工兵出手就好办了

我下个黄茅尖的数据看看，然后两周后去实地体验下三种数据的区别

工兵pro • 2011-02-08 02:04

黄茅尖山头比较平缓，SRTM也足以分辨出主要几个山头，同样的情况适用于三尖，七尖
可以用磨心尖对比一下，无论SRTM还是ASTER都无法分辨从磨心尖到白石尖一连大约10个小山头，这也是我基本摒弃ASTER的理由

Dyson(丁丁) 工兵pro • 2011-02-08 03:39

GLS 2005 - Islands (EO-1)

這是什麼數據？

海島的？

酱铺工兵pro • 2011-02-10 16:35

以个人的有限经验
ASTER的错误率较低，精度似乎要比SRTM高
地形约复杂，越明显
可以试试这两个点

33°58'37.84"N 107°32'58.06"E [srtm经常发生空洞，无法准确的弥补]

或

33°53'27.69"N 108°43'43.03"E [srtm无法分辨较浅的沟谷]

建议一下。
找一个GE的高清山区，用几个不同数据生成矢量等高线
设成不同颜色的线条，叠合在一起，
跟实际地貌图片比较一下，或许比较有说服力。

招财 • 2011-02-08 03:01

两个数据能合并吗?,理论上应该可行吧?

工兵pro • 2011-02-08 03:51

海岛的，使用EO-1卫星的近红外传感器

Dyson(丁丁) • 2011-02-08 04:06

初步比较，个人觉得几种DEM数据的精度为： GLS2005 〉SRTM90 〉Aster GDEM。

这是我老家附近地区的地形，我对其颇为熟悉。
GLS2005比SRTM看起来稍有提高，但没有大幅度的提高，Aster DEM有很多“天坑”，让人用起来容易被误导。
同样2.5米的等高距，GLS2005绘出的等高线比SRTM90更细腻。

[/quote]

同感，大夥有了新的數據，就算是地圖愛好者的新年禮物吧

工兵pro • 2011-02-08 04:12

上图是我用SRTM数据生成的5m等高距图像，可以看到，你认为SRTM无法表现的地形细节这里都有
我这里设定的等高容差是0，线很粗是因为一条线就是一个高度，起伏极小的地形，线就很粗，而悬崖之类的地形则会很细
我觉得你用的参考地形不是太合适，太平缓了，那么小的等高距，已经超出了这些DEM数据的测量误差了

billzeng OP 工兵pro • 2011-02-08 05:30

请问你这样的图能否直接生成矢量数据,导入GPS？我用2.5米的等高距表现出的细节还没有你5米的丰富
看来SRTM90的精度不比GLS205 差

Dyson(丁丁) • 2011-02-08 04:28

Aerial Photography

InformationcostsAerial Photo Single Frames

InformationAircraft Scanners

InformationcostsAntarctic Single Frames

InformationASAS

InformationDOQ

InformationHigh Resolution Orthoimagery

InformationcostsNAPP

InformationcostsNHAP

InformationSLAR

InformationcostsSpace Acquired Photography

AVHRR

InformationAVHRR Composites

InformationAVHRR Orb Seg

Cal/Val Reference Sites

InformationCalibration Sites

Commercial

InformationIKONOS-2

Declassified Data

InformationcostsDeclass 1 (1996)

InformationcostsDeclass 2 (2002)

Digital Elevation

InformationSRTM

Digital Line Graphs

InformationDLG 1:100k

InformationDLG Large Scale

InformationSRTM Water Body Data

Digital Maps

InformationDRG

InformationNational Atlas

EO-1

InformationEO-1 Ali

InformationEO-1 Hyperion

Forest Carbon Sites

InformationFCT - Borneo Island

InformationFCT - Brazil

InformationFCT - Cameroon

InformationFCT - Colombia

InformationFCT - Dem Rep Congo

InformationFCT - Guyana

InformationFCT - Mexico

InformationFCT - Peru

InformationFCT - Sumatra

InformationFCT - Tanzania

InformationFCT - Tasmania Island

Global Land Survey

InformationGLS 2005

InformationGLS 2005 - Islands (EO-1)

InformationGLS LIMA (Antarctica)

InformationGLS 2000

InformationGLS 1990

InformationGLS 1975

HCMM

InformationHCMM

InformationHCMM Digital Source

Land Cover

InformationNLCD 2001

InformationNLCD 1992 - 2001 Retrofit

InformationNLCD 1992

Landsat Archive

InformationL7 SLC-off (2003-present)

InformationL7 SLC-on (1999-2003)

InformationL4-5 TM

InformationL1-5 MSS

InformationL7 Intl Ground Stations (Search Only)

Landsat Legacy

InformationETM+ Pan Mosaics (1999-2003)

InformationTM Mosaics (1984-1997)

InformationETM+ Pan (1999-2003)

InformationETM+ (1999-2003)

InformationTM (1984-1997)

InformationMSS 1-5 (1972-1987)

InformationSys ETM+ L1G (1999-2003)

InformationNALC Triplicates

InformationNLDC

Landsat MRLC

InformationMRLC/MTBS Radiance

InformationMRLC/MTBS Reflectance

NASA LPDAAC Collections

ASTER Collections

InformationASTER Level 1B

Radar

InformationSIR-C

這麼多數據中，有許多是衛星照片。

地形圖用衛星照片+GLS2005+Garmin數據，是目前最為理想的

衛星照片選哪個呢？太多了，看得眼都花了

這是其中之一的喜馬拉雅山脈衛星圖（局部，無坐標）

工兵pro • 2011-02-08 04:42

除了SRTM以外的DEM都是双卫星照片合成得到的，目前能拿到的精度最好的恐怕还是得数google earth买下的digital globle家的, 也只有他家在提供全球1m间隔DEM----因为它照片的采样间隔都0.6m以下了

GLS2005用的的 Landsat 7 卫星的ETM+传感器光学采样间隔就28m了，没得比

Dyson(丁丁) 工兵pro • 2011-02-08 05:05

免得的數據，你還想咋樣？

工兵pro • 2011-02-08 05:41

我自己的山寨产品，无矢量数据，我正在做android版，准备在Defy上用
GLS2005我估计还是值得挖掘的，等我下载了再验证一下

weyawu • 2011-02-08 08:15

如果用GM处理GLS2005的DEM数据，生成等高线，确实如billzeng大侠所言，三种DEM数据中GLS2005的精度最好。但看了工兵大侠用SRTM生成的效果，确实非常棒，如果用工兵大侠的软件算法处理gls2005数据，再能生成矢量数据的话，就能造福广大山友和地图爱好者了。期盼工兵大侠出手，弄一个处理GLS2005生成矢量数据的软件呀！

招财 • 2011-02-10 04:16

大家看看有没有自己那片的

http://www.lvye.org/modules/lvyebb/viewtopic.php?&id=31&post_id=43990366&view=1

Dyson(丁丁) • 2011-02-10 09:41

从地图中可以得知，GLS2005不是很全，只能说部分数据可以下载的到